Cook, T.D., and Campbell, D.T. Quasi-experimentation: Design and analysis issues for field settings. Rand McNally (1979).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....we created five instances of each program; each instance was seeded with only one fault to avoid masking effects. We only employed faults that caused run time failures on the target applications. 3.5. Validity Evaluation Four types of threats can compromise the validity of an experiment [8]. We will now address the possible threats that can affect the outcome of our experiment and limit generalization of our results. The first threat has to do with the construct validity. The measured dependent variables may not be an appropriate reflection of the aspects we want to evaluate. For ....

T.D.Cook and D.T. Cambell, Quasi-ExperimentationDesign and Analysis Issues for Field Settings, Houghton Mifflin Company, 1979.

....Analysis Cruise Requirements Analysis Domain CruiseCar Fig. 2. Parts of form for recording where changes are likely, i.e. change points 3.3 Threats to Validity The validity of the findings is highly dependent on how well threats have been handled. Four types of validity threats [9] are analysed: Threats to conclusion validity, construct validity, internal validity, and external validity. The relationships between these are illustrated in Figure 2. Conclusion validity. Conclusion validity (marked 1 in Figure 2) concerns the relationship between the treatment and the ....

Cook, T.D., Campbell, D.T. Quasi-Experimentation - Design and Analysis Issues for Field Settings. Houghton Mifflin Company. 1979

....than a traditional form of reading . Hence, different conditions need to be studied empirically to identify the techniques strengths and weaknesses. The level of control imposed on the conditions determines whether a study qualifies as a case study, quasi experiment, or controlled experiment [ 7]. The results of the studies help researchers gain an understanding of how a technique works and why the technique is useful. Practitioners, on the other hand, benefit from those studies because the results help them assess the leverage they can expect from a particular technique. This may ....

T.D. Cook and D.T. Campbell. QuasiExperimentation: Design and Analysis Issues for Field Settings. Rand McNally College Publishing Company, Chicago, 1979.

....mortality relationship than those used in previous studies. First, the differential changes in TSPs pollution across counties in this very narrow time frame are both dramatic and abrupt. Consequently, this study has the feature of an interrupted time series design with many control groups (Cook and Campbell 1979). Second, while there is substantial variation across counties in changes in TSPs pollution from 1980 82, we find little evidence of confounding changes in any of the unprecedented number of factors that we can control for. Also, presuming that pollution should have no effect on infant deaths ....

....in the key period. There is unobserved secular factors that affect all types of infant death. Using the language of quasi experimentation , our research design has the feature of an interrupted time series design with both nonequivalent control groups and nonequivalent dependent variables (see Cook and Campbell 1979). 27 For example, Sexton and Hebel (1984) and Evans and Ringel (1999) provide experimental and quasiexperimental evidence, respectively, that maternal smoking reduces infant birth weight by up to 430 550 grams. One potential mechanism is that smoking slows fetal growth by depriving the fetus ....

Cook, Thomas D., and Donald T. Campbell. 1979. Quasi-Experimentation: Design and Analysis Issues for Field Settings. Boston, MA: Houghton Mifflin.

....three strategies that may be followed: a) formal experiments, b) case studies and (c) surveys. In the same the authors express a strong preference for case studies in conducting evaluations in industrial settings. However, another evaluation strategy may be followed, namely quasi experiments (Cook and Campbell 1979). In quasi experimental designs, it is recognized that there are many variables that are difficult to control (as one would attempt to do in a formal experiment) In particular, subjects are not assigned to treatment and control groups randomly, resulting in nonequivalent groups. To illustrate ....

....allow a better level of control of the factors under study; and they can be more easily replicated in different environments. However, this is because it may be difficult or impossible to apply such an approach that social scientists have developed techniques such as quasi experimentation (Cook and Campbell, 1979). Since it takes time for a technology to mature in an organization, case studies are in general also necessary in order to tailor it to the organization based on the feedback from developers and managers [Kitchenham et al. 1995] For example, at the SEL, only after the execution of some case ....

T. Cook and D. Campbell (1979). Quasi-experimentation: Design and Analysis Issues for Field Settings, Rand McNally, Chicago.

....specified in the research hypothesis. Judd et al. add another class of evaluation criteria, construct validity [25] which defines the extent to which the variables successfully measure the theoretical constructs in the hypotheses. Finally, Cook and Campbell add one more class, conclusion validity [18], which defines the extent to which conclusions are statistically valid. The difficulties intrinsic to software engineering (lack of data points, human factors) make it less likely that the validity types can all be satisfied at the same time: e.g. making a study more realistic to achieve a high ....

T. D. Cook, and D. T. Campbell, Quasi-Experimentation: Design and Analysis Issues for field settings, Boston: Houghton Mifflin Co, 1979.

.... as a method of experimental control because it is assumed that over large numbers of subjects uncontrolled factors are distributed evenly over the treatment conditions [87] Since random assignment of subjects to groups is the defining contrast between true experiments and quasi experiments [16] or observational studies [12] this qualifies our study as a quasi experiment. Quasi experiments attempt to preserve as many of the properties of true experimentation as possible given the constraints of the industrial research setting [80] 12 . 11 In a repeated measures design, subjects ....

T. Cook and D. Campbell. Quasi-Experimentation: Design and Analysis Issues for Field Settings. Rand McNally College Publishing Company, Chicago, 1979.

....that each individual method makes trade offs between the fundamental problems they want to avoid and the problems they are willing to accept. Figure 12 (Source: Runkel and McGrath, 1972) shows 8 different research strategies, each of which forces a trade off among different threats to validity (Cook Campbell, 1979). They state that, all other things being equal, it is always best to maximize: generalizability with respect to populations (A) precision in control and measurement of variables related to the behavior of interest (B) and existential realism for the participants and of the context in which ....

Cook, T. and Campbell, D. (1979) QuasiExperimentation: Design and Analysis Issues for Field Settings. Boston, MA: Houghton-Mifflin Company.

No context found.

Cook, T.D., and Campbell, D.T. Quasi-experimentation: Design and analysis issues for field settings. Rand McNally (1979).

No context found.

Cook, T. and D. Campbell (1979). Quasi-Experimentation: Design and Analysis Issues for Field Settings. Rand McNally College Publishing Company.

No context found.

T.D. Cook, D.T. Campbell, Quasi-Experimentation: Design and Analysis Issues in Field Settings, Rand McNally, Chicago, 1979.

No context found.

Cook, Thomas D. & Campbell, Donald T. Quasi-Experimentation: Design and Analysis Issues for Field Settings. Chicago: Rand McNally, 1979.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC